Uncertainty and equifinality in calibrating distributed roughness coefficients in a flood propagation model with limited data

Monte-Carlo simulations of a two-dimensional finite element model of a floo d in the southern paa of Sicily were used to explore the parameter space of distributed bed-roughness coefficients. For many real-world events specifi c data are extremely limited so that there is not only fuzziness in the inf ormation available to calibrate the model, but fuzziness in the degree of a cceptability of model predictions based upon the different parameter values , owing to model structural errors. Here the GLUE procedure is used to comp are model predictions and observations for a certain event, coupled with bo th a fuzzy-rule-based calibration, and a calibration technique based upon n ormal and heteroscedastic distributions of the predicted residuals. The fuz zy-rule-based calibration is suited to an event of this kind, where the inf ormation about the flood is highly uncertain and arises from several differ ent types of observation. The likelihood (relative possibility) distributio ns predicted by the two calibration techniques are similar, although the fu zzy approach enabled us to constrain the parameter distributions more usefu lly, to lie within a range which was consistent with the modellers' a prior i knowledge of the system. (C) 1998 Elsevier Science Limited. All rights re served.